Traditional Filmmaking, AI-Enhanced Post & Marketing
Experimental short film that screened at international festivals
- → Directed and produced the film traditionally on set, with AI tools used exclusively in post-production and marketing
- → Created the poster in Photoshop, Illustrator, and DaVinci, with Midjourney for concept exploration
- → Produced 50+ marketing assets for print, web, and social using Illustrator and GenAI tools, cutting down the creative marketing process significantly
- → Won 2 awards and 1 nomination at 3 festivals
- → Raised $60K with the pitch deck and visuals
Pea Pod Earbuds: AI Product Content Pipeline
Full product shoot generated entirely in Midjourney
I wanted to see if I could produce a complete set of e-commerce product images using only GenAI. Starting from a single product concept (wireless earbuds shaped like peas), I generated listing shots, lifestyle contexts, hero angles, and detail views across multiple prompt iterations.
The biggest challenge was consistency. Each generation produces slight variations in shape, color, and case design. I used low stylization (--s 50) and --style raw to keep outputs grounded and catalog-ready, and iterated on prompts to get closer to a cohesive product identity across all shots.
- → Generated 4 shot types from a single product concept: listing, hero, detail, and lifestyle
- → Used --style raw and low stylization to keep outputs looking like real product photography
- → Ran multiple iterations per shot, refining prompts for realism and consistency
- → Demonstrates a scalable approach to catalog content without physical prototypes or studio time
My First ComfyUI Generation
This camera was one of my first attempts in ComfyUI. Even after multiple iterations, the output still looks rough and low quality. But it showed real promise. With a better GPU, more steps, and refined prompts, I could see that this kind of generation had the potential to reach a whole new level. That was the moment I decided to go all in.
